Reading through recent threads on the 7591 in various classic amps, I'm looking for some opinions on the best way to use it. I just got a quad of EHs. I was thinking UL PP, 6.6K, cathode bias, but almost all the circuits I've seen on the NET use fixed bias. Any reason for this? For information, here are some amps using it.
Eico ST40 ST70
Heathkit AA-100
McIntosh MA 230 integrated
Scott 299C and 299D
Scott LK-72 Integrated
Fisher Model 800-B AM/FM Stereo Receiver (1962)
Fisher Model 500c
Fisher X101
Fisher X202
Sansui 1000A used the 7591 with a 6.6K OPT
Ampeg Amplifier Model G-12
Eico ST40 ST70
Heathkit AA-100
McIntosh MA 230 integrated
Scott 299C and 299D
Scott LK-72 Integrated
Fisher Model 800-B AM/FM Stereo Receiver (1962)
Fisher Model 500c
Fisher X101
Fisher X202
Sansui 1000A used the 7591 with a 6.6K OPT
Ampeg Amplifier Model G-12
Besides what Tom said, the drive requirements of the 7591 are so moderate that there's not much penalty in raising the bias voltage. One other possibility is to use a low-loading scheme; these don't measure well on continuous test tones, but on music, they work quite well and have the advantages of cathode bias's simplicity.
One other possibility is to use a low-loading scheme; these don't measure well on continuous test tones, but on music, they work quite well and have the advantages of cathode bias's simplicity.>
Hi SY - thanks for the post - can you explain more about 'low loading'? I don't know the phrase, and why it doesn't measure well on test tones?
Basically, I don't need power (10 - 15w is OK), I need low distortion. B+ is round 300v. I was thinking of 10v bias, 60mA with a 160 ohm cathode resistor as a start, but I don't know if this would give the best distortion performance. Ideas?
Hi SY - thanks for the post - can you explain more about 'low loading'? I don't know the phrase, and why it doesn't measure well on test tones?
Basically, I don't need power (10 - 15w is OK), I need low distortion. B+ is round 300v. I was thinking of 10v bias, 60mA with a 160 ohm cathode resistor as a start, but I don't know if this would give the best distortion performance. Ideas?
The 7591EH tubes sound great!
IMO they sound better than the original 7591A tubes.
I use them in a couple of refurbished HH Scott amplifiers and in a McIntosh.
You might get some ideas by visiting:
http://hhscott.com/database/vhhs_00011.html
which is part of the www.hhscott.com website.
IMO they sound better than the original 7591A tubes.
I use them in a couple of refurbished HH Scott amplifiers and in a McIntosh.
You might get some ideas by visiting:
http://hhscott.com/database/vhhs_00011.html
which is part of the www.hhscott.com website.
There's an excellent article on "low loading" in Audio; I don't have the reference in front of me, but I'll post it later (pretty sure it was 1959 and an article by Dalzell). In any case, the basic idea is that you use a heavily bypassed cathode resistor for biasing, but you set the anode load to a value more typical of fixed bias arrangements. Then the cathode resistor is sized to provide an idle current fairly close to that of fixed bias instead of the higher idle current for cathode biasing.
Now, when you play music, the bias stays pretty constant because of the bypass cap and because at average signal levels, you're staying pretty much in the class A region. Run a high-power test tone through and the bias voltage starts creeping up, overbiasing the tube and causing crossover distortion.
Myself, I just use a fixed bias and don't worry too much about it. But I've tried the low-loading arrangement and for real music it works as advertised. And it's darn simple.
Now, when you play music, the bias stays pretty constant because of the bypass cap and because at average signal levels, you're staying pretty much in the class A region. Run a high-power test tone through and the bias voltage starts creeping up, overbiasing the tube and causing crossover distortion.
Myself, I just use a fixed bias and don't worry too much about it. But I've tried the low-loading arrangement and for real music it works as advertised. And it's darn simple.
I've never used fixed bias, because my understanding was that cathode bias gave a lower source resistance, which would be very relevant with no global feedback. Crowthorne "Understanding HiFi Circuits" page 31 gives the following effective source resistances
PP UL 1.25
PP triode fixed bias 1.4
PP triode auto bias .4
So according to this, I'll get a lower source resistance in UL than in triode mode with fixed bias. There's no figure for UL fixed bias that I can see - from the text looks as if 'simple ultralinear' is auto bias. You can see my curiosity about fixed and auto bias from this, and I can see why fixed bias wouldn't be an issue with global feedback to pull down the source resitance. Does all this sound correct? I'm going by the textbook here. Andy
PP UL 1.25
PP triode fixed bias 1.4
PP triode auto bias .4
So according to this, I'll get a lower source resistance in UL than in triode mode with fixed bias. There's no figure for UL fixed bias that I can see - from the text looks as if 'simple ultralinear' is auto bias. You can see my curiosity about fixed and auto bias from this, and I can see why fixed bias wouldn't be an issue with global feedback to pull down the source resitance. Does all this sound correct? I'm going by the textbook here. Andy
Guys,
Never loose sight of the fact that the ultralinear topology is a sophisticated form of local NFB.
As was recently pointed out over on AA, pure pentode can be very linear, IF you regulate screen B+ at a fraction of the anode B+. The high Rp of pentode operation (poor damping) is dealt with by using an O/P trafo with a higher reflected impedance. You take advantage of a pentode's ability to swing much closer to the voltage limits than a triode can.
Never loose sight of the fact that the ultralinear topology is a sophisticated form of local NFB.
As was recently pointed out over on AA, pure pentode can be very linear, IF you regulate screen B+ at a fraction of the anode B+. The high Rp of pentode operation (poor damping) is dealt with by using an O/P trafo with a higher reflected impedance. You take advantage of a pentode's ability to swing much closer to the voltage limits than a triode can.
Remember, Crowhurst's examples are for one set of representative operating conditions with one type of tube. Don't read more into them than he intended.
Out of the frying pan, into the fire. Higher turns ratio transformers introduce a new set of problems, like lower power and more distortion from the tube, and more bandwidth problems from the transformer. And you STILL will have a high source impedance unless you use some feedback. Nothing wrong with that, either. If you hate the idea of global feedback for some religious reason, use a local loop from the plates of the o/p tubes back to the cathodes of the driver.
Out of the frying pan, into the fire. Higher turns ratio transformers introduce a new set of problems, like lower power and more distortion from the tube, and more bandwidth problems from the transformer. And you STILL will have a high source impedance unless you use some feedback. Nothing wrong with that, either. If you hate the idea of global feedback for some religious reason, use a local loop from the plates of the o/p tubes back to the cathodes of the driver.
The Fisher X-101 uses 6BQ5/EL84/7189 output tubes. I got a X-101 for free from someone and fixed it. The cathode circuit of the ouput tubes was designed to supply power to the preamp filaments (heaters) and one of the filaments was burned out when I got the amp. I calculated the value of a resistor to substitute for the preamp tube filaments, disconnected the preamp filaments and put in the resistor. Some output tubes would tend to glow red, so I selected four from my collection until I found some that didn't glow when the amp was on. Recently, I tried a new design using a 250 ohm cathode resistor in each channel. One tube glowed a little bit and I played the amplifier for a while. The voltage across each 250 ohm resistor was about 18 V. After that I tried a 500 ohm cathode resistor in each channel and no tube glowed (22 volts across the resistor when amp is warmed up = about 40 milliamps of current through each resistor). I was thinking of trying some resistors between 300 and 400 ohms next. The RCA receiving tube manual lists plate current at around 65-80 mA at various plate voltages for the 6BQ5/EL84 and the 7189.
To be more specific, in the X-101, each output stage is designed to use cathode bias (not fixed bias). I'm trying to redesign my X-101 to use a single cathode resistor (shared between the two output pentodes in each channel) in each channel (along with a cathode resistor bypass capacitor in each channel).
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.